散乱点云的三角网格曲面重建算法

摘要

在采用计算机视觉法获取物体三维重建数据的过程中,为了得到较完整的模型信息,所测得的曲面点通常带有大量冗余,而这些冗余数据的存在大大增加了曲面重建的难度.在此背景下,我们针对散乱无序、无任何几何拓扑信息的密集数据,提出了一种空间三角网格直接剖分算法.该算法能够节省存储空间,提高曲面重建效率,保证输出的曲面网格优质.算法首先对原始数据进行预处理,然后采用空间栅格法及Delaunay空球等准则,扩展动态三角网,最后统一法向量输出完整的三角网格模型.通过实例证明,算法重建速度快,曲面网格质量高.%In order to obtain complete information of surface model in computer vision method,thereare a lot of redundant measured points without any geometric topology,which will make the surface recon-struction more difficult.To solve the problem,an algorithm of triangulation in three dimensions is proposed,which can save data store spaces,increase efficiency and enhance quality of triangulation.This algorithm in-eludes three parts:First,the raw data was preprocessed.Then it was triangulated by means of spatial gridpartition according to delaunay and Circumsphere criterion-Finally,the model will be completed after unit-ing normal vectors.The Experimental results demonstrate that the proposed method may restrict in rapidspeed with high quality grid partition surface.